Policy Teaching through Reward Function Learning